This invention relates to a method for manufacturing a free-standing substrate made of mono-crystalline semiconductor material, in particular a wide bandgap material, and in particular for manufacturing a substrate comprised of gallium nitrate (GaN), aluminum nitrate (AlN) or diamond.
A free-standing substrate is defined as a substrate, whose thickness is sufficient that it carries itself without support. Reasonably, such a substrate must thus have a thickness of at least 100 μm. However, in order to be able to be manipulated in a manufacturing line without risk of breaking, it must generally be thicker.
By way of example, the commercially available free-standing substrates comprised of GaN or AlN have a thickness of 300 μm. Such free-standing substrates comprised of GaN or AlN are used in opto-electronic devices such as LEDs, lasers, sensors or in micro-electronic devices (transistors) or function in a high-temperature environment, or even in the field of hyperfrequency power or power electronics.
A first possibility for obtaining free-standing substrates can comprise fashioning them from a block of the material concerned by sawing and polishing. Unfortunately, at the present time, there is no manufacturing method for GaN or AlN ingots that can be used on an industrial development scale.
The article “Bulk and homoepitaxial GaN growth and characterisation”, Porowski-S, Journal of Crystal Growth, Vol. 189-190, June 1998, pp 153-158, describes a process for growing monocrystalline GaN ingots in the liquid phase under a pressure of 12 to 20 kbars (12 to 20×108 Pa) and at a temperature of between 1400 and 1700° C. These conditions are difficult to implement, however, in the course of mass production. In addition, they only produce crystals having a maximum diameter of 18 mm.
Other teams of researchers have also worked on methods for growing ingots in liquid phase at reduced pressure (less than 2 bars (2×105 Pa)) and at a temperature of 1000° C. The diameter of the crystals obtained is larger, in the vicinity of 50 mm, but the crystalline quality obtained is less satisfactory than in the previously mentioned method.
Finally, the article “Growth and characterisation of GaN single crystals”, Balka et al., Journal of Crystal Growth, Vol 208, January 2000, pp. 100-106, discloses the growth of monocrystalline GaN by sublimation. The manufacturing conditions used are a pressure of less than 1 bar (105 Pa) and a temperature of from 1000 to 1200° C. The crystal quality is very good but the size of the crystal is 3 mm, which is clearly inadequate for the intended applications in the semiconductor field.
At the present time, there is no monocrystalline gallium nitrate or aluminum nitrate on the market, in a massive form, of good quality, having sufficient diameters and at a reasonable price. In order to resolve this problem, one notes in the literature a number of attempts at manufacturing substrates comprised of monocrystalline, free-standing gallium nitrate by thick heteroepitaxy and then eliminating the epitaxy substrate. This thick epitaxy or hydride vapor phase epitaxy (known to the skilled artisan under the acronym HVPE or “hydride vapor phase epitaxy”) consists of producing epitaxial growth of GaN on diverse substrates between 1000° C. and 1100° C. at atmospheric pressure with a view to obtaining a layer of GaN of several tens or hundreds of microns. This technique is advantageous in that it enables one to obtain a good crystal quality and in that it is not necessary to face or to cut the ingots of crude material as in the aforementioned prior art. However, the GaN plates obtained in this fashion have many residual stresses and tensions connected with heteroepitaxy.
Several methods are distinguished according to the nature of the epitaxy support substrate and the technique used to remove the substrate. Thus, according to the article “Physical properties of bulk GaN crystals grown by HVPE”, Melnik et al., MRS Internet Journal of Nitride Semiconductor Research, Vol. 2, Art. 39, a method for growing GaN monocrystals using HVPE on a substrate comprised of monocrystalline silicon carbide (SiC) with removal of the substrate by reactive ionic etching (known to the person skilled in the art under the acronym RIE according to the English expression “reactive ionic etching”). However, removal of this SiC substrate is very time-consuming because it is chemically very inert.
Also, according to the article “Large free-standing GaN substrates by hydride vapor phase epitaxy and laser-induced lift-off,” Kelly et al., Jpn. J. Appl. Phys., Vol 38, 1999, a method for growing GaN by HVPE epitaxy on a sapphire substrate and subsequent removal of the substrate by laser is known (known by the English terminology, “laser-induced lift-off”). Implementing this removal technique is delicate for treating large surfaces because the laser beam scanning is long.
It is also possible to remove the sapphire substrate by mechanical polishing but this method is likewise time-consuming and further presents the risk of breakage of the GaN layer at the time of removal of the substrate that releases the stresses.
In other respects, the article “Preparation of large free-standing GaN substrates by hydride vapor phase epitaxy using GaAs as a starting substrate,” Motoki et al., Jpn. J. Appl. Phys., Vol. 40 (2001), pp. L140-L143 describes a method for growing GaN on a substrate comprised of gallium arsenide (GaAs) by HVPE and then chemical dissolution of the substrate. This technique enables easy removal of the substrate, however, the latter is lost, which is less of an advantage from an economic point of view.
Other attempts have also been made by implementing a technique comprised of growing GaN or aluminum nitride (AlN) on a supporting substrate of silicon (Si {111}) by HVPE and then removing the supporting substrate by chemical etching. This technique has the same drawbacks as those mentioned previously.
Finally, according to U.S. Pat. No. 6,176,925, U.S. patent application Ser. Nos. 2001/0006845 and 2001/00022154 and European patent application 1.045.431, methods are known for obtaining a thick layer of gallium nitride by epitaxial techniques on a seed layer which itself has been obtained by epitaxy. However, none of these four documents mention the possibility of placing a nucleation layer on a support by molecular adhesion bonding.
The important points for realizing free-standing substrates are on the one hand the capacity of realizing thick epitaxy; that is, at least 100 microns while having good crystal quality and on the other hand easy separation of the thick layer from its epitaxy support. The present invention now remedies the aforementioned drawbacks while respecting these important points.